plismbench.models.kaiko_ai module#

Models from Kaiko AI company.

class plismbench.models.kaiko_ai.KaikoViTBase(device: int | list[int] | None = -1, mixed_precision: bool = False)[source]#

Bases: Extractor

Kaiko ViT-Base model available on Pytorch Hub (1-2).

Note

  1. kaiko. ai, Aben, N., de Jong, E. D., Gatopoulos, I., Känzig, N., Karasikov, M., Lagré, A., Moser, R., van Doorn, J., & Tang, F. (2024). Towards large-scale training of pathology foundation models. arXiv. https://arxiv.org/abs/2404.15217

  2. kaiko-ai/towards_large_pathology_fms

Parameters:

  • device (int | list[int] | None = DEFAULT_DEVICE,) – Compute resources to use. If None, will use all available GPUs. If -1, extraction will run on CPU.

  • mixed_precision (bool = True) – Whether to use mixed_precision.

property transform: Compose#

Transform method to apply element wise.

class plismbench.models.kaiko_ai.KaikoViTLarge(device: int | list[int] | None = -1, mixed_precision: bool = False)[source]#

Bases: Extractor

Kaiko ViT-Large model available on Pytorch Hub (1-2).

Note

  1. kaiko. ai, Aben, N., de Jong, E. D., Gatopoulos, I., Känzig, N., Karasikov, M., Lagré, A., Moser, R., van Doorn, J., & Tang, F. (2024). Towards large-scale training of pathology foundation models. arXiv. https://arxiv.org/abs/2404.15217

  2. kaiko-ai/towards_large_pathology_fms

Parameters:

  • device (int | list[int] | None = DEFAULT_DEVICE,) – Compute resources to use. If None, will use all available GPUs. If -1, extraction will run on CPU.

  • mixed_precision (bool = True) – Whether to use mixed_precision.

property transform: Compose#

Transform method to apply element wise.

class plismbench.models.kaiko_ai.Midnight12k(device: int | list[int] | None = -1, mixed_precision: bool = False)[source]#

Bases: Extractor

Midnight-12k model developped by Kaiko AI available on Hugging-Face (1).

Note

  1. https://huggingface.co/kaiko-ai/midnight

Parameters:

  • device (int | list[int] | None = DEFAULT_DEVICE,) – Compute resources to use. If None, will use all available GPUs. If -1, extraction will run on CPU.

  • mixed_precision (bool = True) – Whether to use mixed_precision.

property transform: Compose#

Transform method to apply element wise.